strategy

Should AI Plan Your Week? What Autopilot Proposes vs What You Approve

The Orbitable Team·AI & GTM Strategy·24 Jul 2026·5 min read

Let Orbitable's Autopilot propose your weekly marketing sprint, then decide yourself whether it runs. Autopilot drafts one focus and three to five missions with a why-now for each, plus a carry-over list and an explicit not-doing list, but it proposes only, nothing executes until you approve the plan in the Dock.

What Autopilot actually proposes

Autopilot reads the same world model every agent in your fleet shares, your ICP, brand voice, products, competitors, and uploaded knowledge base, and drafts a weekly plan built around one focus. That focus gets broken into three to five missions, each carrying a why-now so you can see the reasoning rather than just the task. It also lists what carried over from last week's unfinished work, and, separately, what it deliberately chose not to do this week and why. None of it runs. It sits in the Dock as a proposal until you act on it.

Why approval stays with you, not the model

Credits meter completed work. Reviewing a proposal costs nothing, so looking at a plan, editing it, rejecting parts of it, or sitting on it for a day costs you nothing either. That changes the economics of oversight: there's no penalty for scrutiny, only for the work that eventually gets approved and run. Approval happens in the Dock, the same surface where drafts get marked up with pins, sent back, or approved into a Library. Nothing Autopilot proposes touches your credits or your calendar until you say so.

What planning the week yourself looks like instead

Without Autopilot, you're the one deciding which of 10 squads and 50 specialist agents to point at what, this week, out of 22 playbooks, 9 schedule templates, and a 6-phase, 26-step GTM plan. You're also the one remembering what didn't get finished last week and whether that matters, and the one deciding, without writing it down anywhere, what you're consciously skipping. That's not a criticism of manual planning, it's just a description of what disappears when there's no proposal artefact: the carry-over and the trade-off both live in your head instead of in the Dock.

The carry-over list and the not-doing list

These two lists do different jobs. Carry-over stops unfinished missions from quietly vanishing, they reappear in the next week's proposal until you finish them or decide to drop them explicitly. The not-doing list is the more unusual piece: it names what Autopilot is choosing to leave out of this week's plan, and why. Most planning processes, human or automated, hide their trade-offs. This one puts the trade-off in front of you every week, which means you're approving what's excluded as much as what's included.

When to accept the proposal and when to override it

Accept the proposal when it matches what you'd have prioritised anyway and nothing has changed that Autopilot wouldn't know about. It only knows what's in the shared world model, so it has no visibility into this morning's sales call, an informal tip about a competitor's move, or a client escalation that hasn't been logged yet. Override or edit missions when something like that has happened. The Dock is where you make that call, mission by mission, before anything is approved into a running plan.

Autopilot proposal vs planning it yourself

Dimension	Autopilot proposal	Planning it yourself
Starting point	Drafted from the shared world model, one focus, 3 to 5 missions	Starts from a blank calendar and your own memory of last week
Basis for prioritisation	Why-now attached to each mission	Whatever reasoning you hold in your head
Tracking of unfinished work	Explicit carry-over list, reappears until resolved	Depends on whether you wrote it down anywhere
Visibility of trade-offs	Explicit not-doing list, stated up front	Usually implicit, rarely written down
What executes automatically	Nothing, proposal only	Nothing, since you're doing the assigning manually
Where the decision happens	Dock, mission by mission, before anything runs	Wherever you plan, with no shared approval surface
Cost of reviewing	Free, credits are only spent on approved work	No credit cost either way, but no structured review step

FAQ

Does Autopilot ever start work without my approval?

No. Autopilot proposes only, and every weekly plan waits in the Dock until you approve it. No credits are spent on missions that haven't been approved.

What plans include Autopilot?

Autopilot is a paid-plan feature, so it's included on Founder ($89/mo), Team ($349/mo), Agency ($1,999/mo), and Enterprise. There's no separate purchase required on top of any of those tiers.

Do I need the Dock to approve an Autopilot plan?

Yes. Approval happens in the Dock, the same surface used to mark up drafts, approve work into a Library, and raise new requests. Dock access is included by default on Agency and Enterprise, and available as a $49/mo add-on on Founder and Team.

What happens if I ignore the not-doing list?

Nothing happens automatically. Autopilot won't act on anything in the not-doing list unless you explicitly turn it into an approved mission yourself. The list exists so you can see the trade-off it's already making, not to hide work from you.

Does reviewing an Autopilot proposal cost credits?

No. Credits meter completed work, not review, so reading, editing, and either approving or rejecting a weekly plan in the Dock doesn't cost anything.